Michael Regan—Bloomberg Business
(Bloomberg) — The last official statement from the Federal Reserve’s policy committee was more than 700 words, but all that many players in financial markets cared about were whether two of them, “considerable time,” would be present.
Evan Schnidman says he has a better way. The Harvard University PhD and partner William D. MacMillan started a firm called Prattle Analytics that offers software to digest central bank statements like these in an attempt to gauge the Fed’s “mood.” Within a minute, the program produces a score to describe the central bank’s word choices. Known as the Fed Playbook Sentiment Indicator, it’s meant to offer fund managers buy and sell signals for stocks, Treasuries and currencies.
“The verbiage coming out of central banks really is data in my mind,” Schnidman said in a phone interview. “We’re humans, we’re biased, our memories are short and we’re just not quite capable of keeping on an even keel the way that a machine can.”
Prattle’s software can analyze communications from about nine central banks and Schnidman said they plan to bring the number to 15 by the end of this quarter.
The St. Louis-based firm has analyzed central bank communications and subsequent reactions in the media and financial markets over the last 20 years. Within each official Fed statement, meeting minutes, press release, speech and congressional testimony, the software scores each paragraph, sentence, phrase, word and root word, Schnidman said.
“What that allows us to do is not be biased by our own predilections for good buzzwords or bad buzzwords or more substantial or important phrases, but actually look at what the response has been and what the arc of that response has been,” he said.
Prattle has “several dozen” firms testing the data and a few have signed up as customers. They’re targeting global-macro and algorithmic funds, Schnidman said, and they’ve talked with funds focused on options and futures as well as wealth managers.
“They want the information,” Schnidman said. (Disclosure: he has written for Bloomberg View.) “They may not be trading on this as a direct signal, but this is informationally interesting to them. It’s a way to explain to their clients why they are holding out of the market today.”
Some hedge funds may already have similar proprietary software to analyze Fed statements, according to Ezra Rapoport, director of automated strategy and development for Flammarion Capital Partners. He’s another Harvard grad, and he’s skeptical that firms will be able to base many buy and sell decisions solely on Prattle’s product.
“The primary use of central bank communications analytics is risk mitigation,” said Rapoport, whose programs trade Treasuries and futures. “Shops will widen their spreads or otherwise mitigate risk in response to certain pattern matching in the central bank communication.”
Some aggressive and particularly sophisticated players will take directional risk in response to linguistic pattern models, he said, though they are fairly rudimentary, pre-programmed trades.
The type of clients Prattle could gather are “the kind of places for whom the business model dictates buying ALL data sources/signals regardless of value,” Rapoport said in an e-mail. The good news for Schnidman? “There are more than a few of those,” Rapoport said.
To contact the reporter on this story: Michael P. Regan in New York at firstname.lastname@example.org
To contact the editors responsible for this story: Jeff Sutherland at email@example.com
The Prattle Team,